Logistics weighing safety control system based on face recognition

Abstract

The invention relates to the field of logistics, in particular to a logistics weighing safety control system based on face recognition. The identity of a driver and the service legality need to be confirmed and service control needs to be carried out when a logistics transportation vehicle enters and exits from a factory, is loaded and unloaded and is weighed. When a driver enters and exits a weighing field, it is only required that information recognized by an identity card is consistent during entering and exiting registration, and lawbreakers can reduce the net weight of goods or replace goods carried by the driver by replacing license plates or cheating through the vehicle tare weight, so that illegal benefits are gained. When the vehicle enters and exits the weighing field, information shooting images of the vehicle and a driver of the vehicle are identified and compared, and whether the vehicle appearance, the license plate number, the driver and the identity card registration information of the vehicle registered by swiping the identity card are consistent or not is confirmed; the behavior of illegally replacing the driver or driving the vehicle is warned in time in the operation process, and the follow-up service process is stopped.

Description

Technical Field

[0001] The present invention relates to the field of logistics, and in particular to a logistics weighing safety control system based on face recognition. Background Art

[0002] Logistics weighing refers to the use of truck scales to accurately weigh the goods transported by vehicles during the logistics process, and using the weighing results as an important basis for logistics management, contract execution, cargo safety, etc.

[0003] When logistics vehicles enter and leave the factory, load and unload vehicles, and weigh, they need to confirm the driver's identity and business legitimacy for business control, and use the weighing results as an important basis for logistics management, contract execution, cargo safety, etc. In the current logistics weighing process, freight vehicles weigh cargo at the weighing site, and only need to ensure that the entry and exit registration information is consistent when entering and leaving the weighing site. Criminals can reduce the net weight of the cargo or replace the cargo carried by the driver by changing the license plate or using the vehicle's tare weight to cheat, thereby seeking illegal profits.

[0004] Replacing vehicles and drivers during transportation and operations will also increase the difficulty of logistics management. Companies that transport goods need to re-register, review and track the replaced vehicles to ensure the safety and accurate delivery of the goods. This not only increases the workload, but also increases the risk of errors in the transported goods.

[0005] Therefore, an improved logistics weighing safety control system is needed, which can accurately identify the information of vehicles entering and leaving the weighing and the drivers inside, and provide reminders when abnormalities occur, while retaining the old registration system and safety response measures. Summary of the invention

[0006] In order to achieve the above functions, the present invention provides a logistics weighing safety control system based on face recognition, which includes the following functions:

[0007] A logistics weighing safety control system based on face recognition, characterized in that the logistics weighing safety control system based on face recognition includes:

[0008] Capture module: used to collect images of the vehicle and the driver when the driver swipes his ID card to register, and send the processed images to the recognition module;

[0009] Recognition module: receives the image sent by the acquisition module, identifies the face image and the vehicle image therefrom, extracts the facial feature information and the vehicle information, and sends the facial feature information and the vehicle information to the comparison module;

[0010] Comparison module: query the vehicle registration information recorded in the storage module. When there is no information identical to the currently registered vehicle, the current vehicle information and facial feature information are processed and sent to the storage module for storage. When there is information identical to the currently registered vehicle, the received vehicle information is compared with the facial feature information corresponding to the stored vehicle information, and then the comparison result is sent to the decision module.

[0011] Judgment module: judges the comparison results received and decides whether to allow the vehicle to pass according to the actual situation. If the judgment is not passed, it will be transferred to manual processing and returned to the acquisition module;

[0012] Storage module: used to store the vehicle registration information and receive the vehicle information and facial feature information transmitted by the comparison module.

[0013] Furthermore, the images captured by the acquisition module include the image of the vehicle and the image of the driver in the vehicle, and the format of the images is YUV. The purpose of the image processing is to convert the image in YUV format into an image in RGB format. When the registration information obtained by the vehicle driver by swiping his ID card to register does not exist in the storage module, relevant staff will check whether the registration information is consistent with the image captured by the acquisition module. If they are inconsistent, the vehicle will be prevented from passing and rectification will be carried out.

[0014] Furthermore, the identification method of the identification module is:

[0015] S1. The image sent by the acquisition module is converted into the RGB numerical image;

[0016] S2. converting the RGB numerical image into an RGB matrix;

[0017] S3. Identify the information corresponding to the RGB matrix;

[0018] The generation formula of the RGB matrix is:

[0019]

[0020] Wherein, M represents the RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, the position of the matrix corresponds to the position of the RGB image, c1, c2, c3 are proportional coefficients, the sum of c1, c2, c3 is 1, R is the brightness of red at a certain position in the RGB image, G is the brightness of green at a certain position in the RGB image, B is the brightness of blue at a certain position in the RGB image, and c4 is a normalization coefficient.

[0021] Furthermore, the vehicle information includes a license plate number and a vehicle appearance. The recognition module identifies the license plate and the facial image of the driver of the vehicle from the image sent by the acquisition module and marks the position in the image sent by the acquisition module. According to the recognition method of the recognition module, a license plate number RGB matrix corresponding to the license plate number, a facial image RGB matrix corresponding to the facial image, and a vehicle appearance RGB matrix corresponding to the vehicle appearance are generated.

[0022] Furthermore, the method for comparing the vehicle appearance in the vehicle information is that the comparison module obtains the vehicle appearance similarity according to the difference of the vehicle appearance RGB matrix of the vehicles participating in the comparison, and the calculation formula of the vehicle appearance similarity is:

[0023]

[0024] Where W is the vehicle appearance similarity, n and p are the total number of rows and columns of the vehicle appearance RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, M1 1ij M1 represents the RGB value of the vehicle appearance RGB matrix in the i-th row and j-th column of the first party participating in the comparison, 2ij Represents the RGB value of the RGB matrix of the second party participating in the comparison at the i-th row and j-th column.

[0025] Furthermore, the license plate number RGB matrix corresponding to the license plate number can be compared with a pre-established character matrix library to obtain characters corresponding to the license plate number RGB matrix, and the license plate number can be obtained by integrating the obtained characters; the facial feature information is a curve set of the RGB values ​​obtained according to the corresponding RGB matrix of the face image, which varies with position. The comparison module determines the similarity of the facial feature information of the two parties to the comparison according to the degree of deviation of the curve sets of the two parties to the comparison. The calculation formula of the facial feature information similarity is:

[0026]

[0027] Where F is the similarity of the facial feature information, n and p are the total number of rows and columns of the RGB matrix of the face image, i represents the i-th row in the matrix, j represents the j-th column in the matrix, and M2 1ij M2 represents the RGB value of the RGB matrix of the face image of the first party participating in the comparison in the i-th row and j-th column, 2ij represents the RGB value of the RGB matrix of the face image of the second party participating in the comparison in the i-th row and j-th column, f represents the error correction parameter, and m represents the convergence coefficient.

[0028] Furthermore, the decision formula of the decision module is:

[0029]

[0030] Among them, T is the judgment threshold, Z represents whether the license plate number information of the two parties being compared is consistent, α1 and α2 are normalization parameters, W is the vehicle appearance similarity, and F is the facial feature information similarity; when the license plate number information of the two parties being compared is completely consistent, the size of Z is 1, and when the license plate number information of the two parties being compared is not completely consistent, the size of Z is 0. When T is greater than 0.5, the judgment module judges that the information of the two parties being compared is the same, and when T is less than 0.5, the judgment module judges that the information of the two parties being compared is different, and the judgment module issues a warning to the relevant personnel.

[0031] Furthermore, the acquisition module adjusts its position according to different vehicles to capture images of the same position of the vehicles, the images including the front image and the side image of the vehicle, the images also including the facial image of the driver of the vehicle, the front image of the license plate, and the front image and the side image of the vehicle's appearance.

[0032] Furthermore, the recognition module fails to recognize the vehicle information of the vehicle and the facial feature information of the driver in the vehicle from the image sent by the acquisition module, and cannot mark the facial image and the license plate, indicating that the acquisition module has not captured a usable image. The recognition module sends a command to the acquisition module to enable the acquisition module to re-capture the image until the recognition module can recognize the vehicle information of the vehicle and the facial feature information of the driver in the vehicle.

[0033] Furthermore, when the judgment module fails to pass the judgment, if the reason for the failure of the judgment is that the vehicle information and the driver in the vehicle have changed due to special reasons, relevant personnel can modify the vehicle information corresponding to the registration information and the facial feature information of the driver in the vehicle stored in the storage module.

[0034] Beneficial effects:

[0035] 1. Identify and compare the images of vehicles and their drivers when they enter and leave the weighing site, and confirm whether the vehicle appearance, license plate number, and driver identified by the camera are consistent with the ID card information of the vehicle registered by swiping the ID card, so that the illegal replacement of drivers or driving of vehicles will be promptly warned during the operation and the subsequent business process will be terminated, making up for the original defect of whether the ID card registration information is consistent with the actual transport vehicle and driver;

[0036] 2. The images used for identification are collected at multiple angles through cameras, ensuring sufficient collection of vehicle and driver feature information and improving the reliability of image recognition. The collected images are converted into RGB data before image recognition, which improves the efficiency of the image recognition program and its applicability to different recognition devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 The present invention is a flowchart of the working process when a vehicle enters a parking lot;

[0038] Figure 2 The working flow diagram of the present invention when the vehicle exits the vehicle;

[0039] Figure 3 The overall work flow chart of information comparison for the present invention;

[0040] Figure 4 The workflow diagram of image recognition for the present invention;

[0041] Figure 5 It is a composition structure diagram of the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0043] The present invention provides a logistics weighing safety control system based on face recognition, which includes the following contents:

[0044] like Figure 5 As shown, the present invention includes the following modules:

[0045] Capture module: used to collect images of the vehicle and the driver when the driver swipes his ID card to register, and send the processed images to the recognition module;

[0046] Recognition module: receives the image sent by the acquisition module, identifies the face image and the vehicle image therefrom, extracts the facial feature information and the vehicle information, and sends the facial feature information and the vehicle information to the comparison module;

[0047] Comparison module: query the vehicle registration information recorded in the storage module. If there is no information identical to the currently registered vehicle, the current vehicle information and facial feature information are processed and sent to the storage module for storage. If there is information identical to the currently registered vehicle, the received vehicle information is compared with the facial feature information corresponding to the stored vehicle information, and then the comparison result is sent to the decision module.

[0048] Judgment module: judges the comparison results received and decides whether to allow the vehicle to pass according to the actual situation. If the judgment is not passed, it will be transferred to manual processing and returned to the acquisition module;

[0049] Storage module: used to store the vehicle registration information and receive the vehicle information and facial feature information transmitted by the comparison module.

[0050] like Figure 1-Figure 3 As shown, vehicles transporting goods need to enter the weighing site for weighing in order to calculate the subsequent logistics costs and ensure transportation safety. Before entering the site, the driver needs to swipe his ID card on the identification device to register. The information related to the ID card includes vehicle information, cargo information, cargo transaction information, time information, security information and driver information. The cargo information and cargo transaction information will be confirmed during the weighing process. After the weighing is completed, the driver of the vehicle transporting the goods needs to register his ID card again to update the time information and security information before leaving the site.

[0051] The information related to the ID card cannot be changed at will. The verification of the information related to the ID card is the responsibility of the weighing-related equipment. The driver and vehicle holding the ID card may be changed during the weighing process. The purpose of this logistics weighing safety control system based on face recognition is to ensure that the vehicles and their drivers entering and leaving the weighing site through ID card registration and identification should be consistent with the information corresponding to the recorded ID card and there should be no obvious changes when entering and leaving the site.

[0052] The acquisition module is a multi-angle camera, which is used to collect multi-angle vehicle and personnel images when the vehicle transporting goods is registered to enter and exit the weighing site. These images include the front view, side view and back view of the vehicle transporting goods, and the front image of the vehicle's driving position when registering the ID card. The front image of the vehicle transporting goods covers the front license plate and the front image of the vehicle's driving position, the side image covers the upper half of the vehicle, and the back image covers the front image of the rear license plate. The front image during ID card registration covers the front image of the driver's face.

[0053] The same vehicle enters and exits the weighing site in opposite directions, so the acquisition module has multiple cameras in opposite directions to alternately acquire images to ensure that the acquired images are independent of the vehicle's direction.

[0054] The images collected by the camera used by the acquisition module are all in YUV format. The YUV format occupies very low bandwidth, which is conducive to the camera taking a large number of images in a short time. However, this format is incompatible with many common electronic devices, and the YUV format is difficult to quantize, which is not convenient for subsequent image processing. Therefore, the acquisition module will process the YUV format image into RGB format before transmitting the image. RGB format is the most compatible color format, and color data can be easily divided into three colors: red, green, and blue, which is convenient for subsequent quantization processing. The formula for converting YUV format to RGB format is:

[0055] R=Y+1.402\(V-128)

[0056] G=Y-0.364\(U-128)

[0057] B=Y+1.772\(U-128)

[0058] The R, G, and B in the formula correspond to the proportions of red, green, and blue contained in the pixels in the image, respectively. Y represents brightness, that is, the grayscale value. U and V represent chromaticity, which is used to describe the color and saturation of the image and is used to specify the color of the pixel. The decimal in the formula is not a fixed value and is affected by factors such as the brightness of the actual environment. The acquisition module will automatically adjust according to changes in the environment.

[0059] During the image collection process, if the information corresponding to the ID card is not recorded in the logistics weighing safety control system, the relevant staff will take different measures according to the actual situation: if the registration purpose of the vehicle whose image is collected at this time is to leave the weighing site, it means that the information corresponding to the ID card is wrong, and the relevant personnel prohibit the vehicle from leaving the site and report to the superior for processing; if the registration purpose of the vehicle whose image is collected is to enter the weighing site, it means that the vehicle is entering the weighing site for the first time. The recognition module will use the image collected at this time as the basis for future judgments. The relevant personnel will visually check the vehicle at this time to ensure that the relevant information of the vehicle and the driver is consistent with the information corresponding to the ID card, and then associate the information corresponding to the ID card with the relevant information of the vehicle driver. If there is inconsistency, the vehicle should be stopped from passing and rectification should be carried out.

[0060] like Figure 4As shown, after the acquisition module processes the color format of the image, the acquisition module sends the processed image to the recognition module. The recognition module needs to identify the license plate number, the vehicle appearance and the driver's facial image from the received image, wherein the license plate number and the driver's facial image require high clarity for subsequent comparison. The recognition module can identify the position of a specific image in the image after big data learning. The recognition module needs to identify the license plate number and the driver's face position from the image sent by the acquisition module and mark them. If the recognition module does not identify the vehicle information of the vehicle and the facial feature information of the driver in the vehicle from the image sent by the acquisition module, that is, it cannot mark the facial image and the license plate, indicating that the image collected by the acquisition module is unavailable, the recognition module sends a command to the acquisition module to enable the acquisition module to re-acquire the image until the recognition module can identify the vehicle information of the vehicle and the facial feature information of the driver in the vehicle. In this process, the acquisition module will continuously adjust the position angle until the recognition module successfully recognizes it. After the recognition module successfully recognizes it, the acquisition module will record the current position and angle as a reference for subsequent adjustments.

[0061] The identification method of the identification module is:

[0062] S1. Convert the image sent by the acquisition module into an RGB numerical image. The aforementioned RGB images are still different images. In order to facilitate the processing of the image, it is necessary to digitize the pixels in the image and replace each pixel in the image with its corresponding R, G, B values, thereby converting the image into an RGB numerical image.

[0063] S2. Convert the RGB numerical image into an RGB matrix, perform weighted summing and normalization of the R, G, and B values ​​corresponding to each pixel in a certain ratio to obtain the RGB value corresponding to each pixel, and arrange the order of the RGB values ​​corresponding to the pixels to obtain the RGB matrix.

[0064] The formula for generating the RGB matrix is:

[0065]

[0066] Among them, M represents the RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, the position of the matrix corresponds to the position of the RGB image, c1, c2, and c3 are proportional coefficients, the sum of c1, c2, and c3 is 1, R is the brightness of red at a certain position in the image, G is the brightness of green at a certain position in the image, B is the brightness of blue at a certain position in the image, and c4 is the normalization coefficient.

[0067] The scale factor is dynamically fine-tuned according to changes in the image shooting environment. The purpose of the recognition module to process the image is to perform image recognition. It does not need to accurately extract and restore the image itself. Therefore, the recognition step can be appropriately simplified, and the normalization coefficient is directly changed linearly with the change of the scale factor, so that the RGB value is maintained at the level of 0 to 1 for subsequent calculations.

[0068] S3. Identify the information corresponding to the RGB matrix. The information to be identified in the RGB matrix includes vehicle information and driver's facial feature information. The vehicle information includes the license plate number and vehicle appearance: the recognition module identifies the position of the license plate and the driver's face image in the RGB matrix according to the position marked in the image sent by the acquisition module, and uses the RGB matrix as a whole as the RGB value corresponding to the vehicle appearance. The recognition module extracts the features of the license plate number, vehicle appearance and face image as follows:

[0069] For the license plate number, the vehicle has two license plates, one in front and one at the back. The recognition module can mark the positions of the two license plates in the image and extract the corresponding positions of the two license plates in the RGB matrix, compare the corresponding positions of the two license plates in the RGB matrix, retain the same parts of the two, and finally obtain the partial RGB matrix closest to the actual license plate number. The partial RGB matrix is ​​located at the position marked as the license plate number in the RGB matrix. The recognition module is pre-set with a character matrix library generated through a large amount of training. The character matrices in the character matrix library are all common license plate characters obtained through a large amount of pre-training, including the abbreviations, letters and numbers of various provinces. By comparing the partial RGB matrix with the characters in the character matrix library one by one, the character set corresponding to the partial RGB matrix can be obtained, and the character set is the license plate number.

[0070] For the vehicle appearance, the vehicle appearance refers to the shape and color of the vehicle. The purpose of the recognition module is comparison rather than precise recognition. It does not need to accurately extract the shape and color changes of the vehicle. It only needs to collect enough features in the vehicle appearance. Therefore, the vehicle appearance directly uses the entire RGB matrix for comparison.

[0071] For face images, compared with license plate numbers, the changes in faces are more subtle and complex. For the sake of convenience, the acquisition module can only capture faces at a long distance, so the features on the face image cannot be directly extracted. Considering that the recognition module has divided the area belonging to the face, only the features that the face can show are considered. The curve of the change of RGB values ​​in the RGB matrix area belonging to the face with the change of position is used as the basis for face comparison.

[0072] The comparison module receives the RGB matrix processed by the recognition module and compares the RGB matrix of the same category with the same information registered using the ID card. The two sides of the comparison are images taken when the vehicle enters and exits the site. The corresponding vehicle and driver are the same under normal circumstances. The positions of the license plate number, vehicle appearance and face image in the image collected by the acquisition module are also the same, so the RGB matrix processed by the recognition module can be directly compared. The comparison principle is:

[0073] Comparison of the license plate numbers in the vehicle information: directly compare the character sets obtained by the recognition modules of both parties. The recognition accuracy of the license plate numbers is high, so there are only two results: the same or different.

[0074] Comparison of vehicle appearance in vehicle information: The recognition accuracy of vehicle appearance is low and it is easy to undergo subtle changes. Therefore, when the two compared parties are the same vehicle, the comparison results will not be completely equal.

[0075] The vehicle appearance similarity is introduced during the comparison to represent the degree of similarity between the two sides of the comparison. The vehicle appearance similarity is obtained by the comparison module based on the difference in the RGB matrix of the vehicle appearance of the two sides of the comparison. The calculation formula for the vehicle appearance similarity is:

[0076]

[0077] Where W is the vehicle appearance similarity, n and p are the total number of rows and columns of the vehicle appearance RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, and M1 1ij M1 represents the RGB value of the first-party vehicle appearance RGB matrix in the i-th row and j-th column, 2ij Represents the RGB value of the second party RGB matrix participating in the comparison at the i-th row and j-th column.

[0078] For the comparison of face images: the curve obtained by the recognition module is the feature of the face image. Using all the curves as the features of the face image at the same time can more accurately and completely reflect the features of the face. Therefore, the facial feature information is introduced. The facial feature information is a set of curves of RGB changing with position obtained according to the corresponding face image RGB matrix. The comparison module determines the similarity of the facial feature information of the two parties according to the deviation degree of the curve sets of the two parties. The calculation formula of the facial feature information similarity is:

[0079]

[0080] Among them, F is the similarity of facial feature information, n and p are the total number of rows and columns of the RGB matrix of the face image, i represents the i-th row in the matrix, j represents the j-th column in the matrix, and M2 1ijM2 represents the RGB value of the RGB matrix of the first party’s face image in the i-th row and j-th column. 2ij represents the RGB value of the RGB matrix of the second face image participating in the comparison in the i-th row and j-th column, f represents the error correction parameter, and m represents the convergence coefficient.

[0081] When a vehicle leaves the scene, the judgment module determines whether the vehicle and the driver are inconsistent based on the comparison results of the comparison module, and decides whether to allow the vehicle to pass. The data involved in the judgment include the comparison results of the license plate number, vehicle appearance and facial features received by the judgment module from the comparison module. The recognition accuracy of the license plate number is relatively high. If the character set corresponding to the RGB number is inconsistent, it means that the vehicle's license plate number is inconsistent. There is a certain error in the vehicle's appearance, so the vehicle's appearance is inconsistent only when the vehicle's appearance similarity is lower than a certain value. Different faces have a certain amount of the same features. At the same time, faces are affected by factors such as light and shadow, and the collection angle of the collection module, and there are also certain errors. The common features shared by different faces and the features with errors caused by other factors on the same face account for a very low proportion of all facial features, so the similarity of facial feature information conforms to the exponential distribution.

[0082] Based on the characteristics and practical importance of the character set corresponding to the license plate number, the vehicle appearance similarity, and the facial feature information similarity, the decision formula of the decision module can be obtained as follows:

[0083]

[0084] Among them, T is the decision threshold, Z represents whether the license plate information of the two parties is consistent, α1 and α2 are normalization parameters, W is the vehicle appearance similarity, and F is the facial feature information similarity;

[0085] α1 and α2 are obtained from past experience using big data training, and can be dynamically adjusted by the judgment module according to changes in the environment. When the license plate number information of the two parties is completely consistent, the size of Z is 1. When the license plate number information of the two parties is not completely consistent, the size of Z is 0. When T is greater than 0.5, the judgment module judges that the information of the two parties is the same. When T is less than 0.5, the judgment module judges that the information of the two parties is different, and the judgment module issues a warning to the relevant personnel.

[0086] In addition, only after the ID card registration and the judgment of the judgment module are passed at the same time, can the vehicle enter and exit the weighing site.

[0087] On the other hand, when the judgment module fails the judgment, in addition to problems with the vehicle and the driver, it is also possible that the logistics weighing safety system fails or special circumstances require the driver or freight vehicle to be replaced. At this time, the relevant personnel can manually input instructions to update the vehicle information and facial feature information corresponding to the ID card registration information stored in the storage module.

[0088] The above embodiments only express the preferred implementation modes of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications, improvements and substitutions can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the attached claims.

Claims

1. A logistics weighing safety control system based on face recognition, characterized in that: The logistics weighing safety control system based on face recognition includes: Capture module: used to collect images of the vehicle and the driver when the driver swipes his ID card to register, and send the processed images to the recognition module; Recognition module: receives the image sent by the acquisition module, identifies the face image and the vehicle image therefrom, extracts the facial feature information and the vehicle information, and sends the facial feature information and the vehicle information to the comparison module; Comparison module: query the vehicle registration information recorded in the storage module. When there is no information identical to the currently registered vehicle, the current vehicle information and facial feature information are processed and sent to the storage module for storage. When there is information identical to the currently registered vehicle, the received vehicle information is compared with the facial feature information corresponding to the stored vehicle information, and then the comparison result is sent to the decision module. Judgment module: judges the comparison results received and decides whether to allow the vehicle to pass according to the actual situation. If the judgment is not passed, it will be transferred to manual processing and returned to the acquisition module; Storage module: used to store the vehicle registration information and receive the vehicle information and facial feature information transmitted by the comparison module.

2. According to the face recognition-based logistics weighing safety control system of claim 1, it is characterized in that: The images captured by the acquisition module include the image of the vehicle and the image of the driver in the vehicle, and the format of the images is YUV. The purpose of the image processing is to convert the image in YUV format into an image in RGB format. When the registration information obtained by the vehicle driver by swiping his ID card to register does not exist in the storage module, the relevant staff will check whether the registration information is consistent with the image captured by the acquisition module. If they are inconsistent, the vehicle will be prevented from passing and rectification will be carried out.

3. According to the face recognition-based logistics weighing safety control system of claim 1, it is characterized in that: The identification method of the identification module is: S1. The image sent by the acquisition module is converted into the RGB numerical image; S2. converting the RGB numerical image into an RGB matrix; S3. Identify the information corresponding to the RGB matrix; The generation formula of the RGB matrix is: Wherein, M represents the RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, the position of the matrix corresponds to the position of the RGB image, c1, c2, c3 are proportional coefficients, the sum of c1, c2, c3 is 1, R is the brightness of red at a certain position in the RGB image, G is the brightness of green at a certain position in the RGB image, B is the brightness of blue at a certain position in the RGB image, and c4 is a normalization coefficient.

4. According to the face recognition-based logistics weighing safety control system of claim 1, it is characterized in that: The vehicle information includes a license plate number and a vehicle appearance. The recognition module recognizes the license plate and the facial image of the driver of the vehicle from the image sent by the acquisition module and marks the position in the image sent by the acquisition module. According to the recognition method of the recognition module, a license plate number RGB matrix corresponding to the license plate number, a facial image RGB matrix corresponding to the facial image, and a vehicle appearance RGB matrix corresponding to the vehicle appearance are generated.

5. A logistics weighing safety control system based on face recognition according to claim 4, characterized in that: The method for comparing the vehicle appearance in the vehicle information is that the comparison module obtains the vehicle appearance similarity according to the difference of the vehicle appearance RGB matrix of the vehicles participating in the comparison, and the calculation formula of the vehicle appearance similarity is: Where W is the vehicle appearance similarity, n and p are the total number of rows and columns of the vehicle appearance RGB matrix, i represents the i-th row in the matrix, j represents the j-th column in the matrix, M1 1ij M1 represents the RGB value of the vehicle appearance RGB matrix in the i-th row and j-th column of the first party participating in the comparison, 2ij Represents the RGB value of the RGB matrix of the second party participating in the comparison at the i-th row and j-th column.

6. A logistics weighing safety control system based on face recognition according to claim 4, characterized in that: The license plate number RGB matrix corresponding to the license plate number can be compared with a pre-established character matrix library to obtain characters corresponding to the license plate number RGB matrix, and the license plate number can be obtained by integrating the obtained characters; the facial feature information is a curve set of the RGB values ​​obtained according to the corresponding RGB matrix of the face image, which varies with the position. The comparison module determines the similarity of the facial feature information of the two parties to the comparison according to the degree of deviation of the curve sets of the two parties to the comparison. The calculation formula of the facial feature information similarity is: Where F is the similarity of the facial feature information, n and p are the total number of rows and columns of the RGB matrix of the face image, i represents the i-th row in the matrix, j represents the j-th column in the matrix, and M2 1ij M2 represents the RGB value of the RGB matrix of the face image of the first party participating in the comparison in the i-th row and j-th column, 2ij represents the RGB value of the RGB matrix of the face image of the second party participating in the comparison in the i-th row and j-th column, f represents the error correction parameter, and m represents the convergence coefficient.

7. A logistics weighing safety control system based on face recognition according to claim 1, characterized in that: The decision formula of the decision module is: Among them, T is the judgment threshold, Z represents whether the license plate number information of the two parties being compared is consistent, α1 and α2 are normalization parameters, W is the vehicle appearance similarity, and F is the facial feature information similarity; when the license plate number information of the two parties being compared is completely consistent, the size of Z is 1, and when the license plate number information of the two parties being compared is not completely consistent, the size of Z is 0. When T is greater than 0.5, the judgment module judges that the information of the two parties being compared is the same, and when T is less than 0.5, the judgment module judges that the information of the two parties being compared is different, and the judgment module issues a warning to the relevant personnel.

8. The logistics weighing safety control system based on face recognition according to claim 1 is characterized in that: The acquisition module adjusts its position according to different vehicles to acquire images of the same position of the vehicles, wherein the images include the front image and the side image of the vehicles, the images also include the facial image of the driver of the vehicles, the front image of the license plate, and the front image and the side image of the vehicle appearance.

9. The logistics weighing safety control system based on face recognition according to claim 1 is characterized in that: The recognition module fails to recognize the vehicle information of the vehicle and the facial feature information of the driver in the vehicle from the image sent by the acquisition module, and cannot mark the facial image and the license plate, indicating that the acquisition module has not acquired a usable image. The recognition module sends a command to the acquisition module to enable the acquisition module to re-acquire the image until the recognition module can recognize the vehicle information of the vehicle and the facial feature information of the driver in the vehicle.

10. The logistics weighing safety control system based on face recognition according to claim 1 is characterized in that: When the judgment module fails the judgment, if the reason for the failure of the judgment is that the vehicle information and the driver in the vehicle have changed due to special reasons, the relevant personnel can modify the vehicle information corresponding to the registration information and the facial feature information of the driver in the vehicle stored in the storage module.

Images